BIG DATA, SMALL CREDIT - The Digital Revolution and Its Impact on Emerging Market Consumers - IssueLab
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
BIG DATA, SMALL CREDIT
The Digital Revolution and Its Impact
on Emerging Market Consumers
By Arjuna Costa, Anamitra Deb, and Michael Kubzansky
TABLE OF CONTENTS
SECTION 1: The Rapidly Evolving Market for Consumer Credit. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Figure 1 – Big Data, Small Credit is at the Intersection of Three Significant Global Trends . . . . . . . . . . . . . . . . . . 5
The Digital Revolution and Financial Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
The BDSC Advantage: Lower Cost, Better Reach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
The ABCDs of Early Business Models: Algorithms, Big Data, Consumers, and Distribution. . . . . . . . . . . . . . . . . . 7
Figure 2 – Multiple Data Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
An Evolving Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
SECTION 2: BDSC Early Adopters: Why and How Do They Borrow?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
The Voice of the Consumer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Early Adopters: Young, Savvy and Educated . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Figure 3 – How Consumers Use Their Mobile Phones Daily . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Figure 4 – The Early Adopters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Borrowing Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Figure 5 – Top Uses for Loans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Figure 6 – Where Participants Applied for Loans. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
SECTION 3: BDSC Early Adopters: Their Evolving Trust Calculus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Trust, Privacy, and Credit: The Bottom Line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Figure 7 – What Data Consumers Consider Private. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Figure 8 – Information Consumers are Willing to Share to Improve
Their Chances of Getting a Loan or a Bigger Loan. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Figure 9 – Top Offenses that Erode Trust in Financial Insitutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Looking Ahead. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
SECTION 4: Delivering on BDSC’s Promise. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
A Call to Action. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Building a Trust-Enhancing Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Figure 11 – What Has a Positive Impact on Consumer Trust in Financial Institutions?. . . . . . . . . . . . . . . . . . . . . 31
Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
ENDNOTES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
ACKNOWLEDGMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2
SE C TIO N 1
THE RAPIDLY EVOLVING MARKET
FOR CONSUMER CREDIT
We are on the verge of massive change. Technological advances in credit assessment are poised to deliver huge
impact by bringing formal, accessible, and affordable credit to hundreds of millions of aspiring middle-class
consumers in emerging markets. At the forefront of this change is a burgeoning new field that we’re calling “Big
Data, Small Credit” (BDSC).
Around the world, many emerging-market consumers remain severely limited in their access to formal financial
services, particularly unsecured credit. In India alone, in 2014, more than 400 million people borrowed money—but
fewer than one in seven were approved for a formal loan.1 Indeed, this experience of being “invisible” to formal
lenders is prevalent among billions of “thin file” or “no file” consumers living in nearly all of today’s emerging markets.
But these consumers may not remain “invisible” to formal lenders for long, thanks in part to their rising use of
technology. Well over 650 million adults in India—four out of every five—already have a mobile phone in their
pocket, and most of these will be smartphones by 2020. More than 240 million Indians have access to the Internet
and social media.2 Seven in 10 users of mobile broadband smartphone in India regularly stream videos on their
phones; six in 10 use social networks.3 And every time these individuals make a phone call, send a text, browse
3
the Internet, engage social media networks, or top up their prepaid cards, they deepen the digital footprints they
are leaving behind.
These digital footprints are helping to spark a new kind of revolution in lending. In the last few years, a cluster of
fast-emerging and innovative firms has begun to use highly predictive technologies and algorithms to interrogate
and generate insights from these footprints. These BDSC firms are using varied forms of nontraditional data—from
mobile call data records and bill payments to Internet browsing patterns and social media behavior—to create a
new way to assess consumer risk, determine the creditworthiness of previously “invisible” consumers, and con-
sequently offer convenient, quicker, and often cheaper loans to the previously underserved. Their prime offering:
unsecured, short-term, small-ticket consumer credit served at a dramatically lower cost than traditional loans.
Currently, in India alone, BDSC firms and their technologies have the ability to vastly expand the number of
individuals who could gain access to formal credit by nearly 100-160 million new consumers.4 In the world’s six
biggest emerging economies—China, Brazil, India, Mexico, Indonesia, and Turkey—BDSC has the potential to
help between 325 million and 580 million people gain access to formal credit for the first time5, creating a market
opportunity that is too big to ignore.
Moreover, if any individual with a basic
feature phone can be assessed based on The conversion of a formerly “invisible”
these digital footprints, then this “credit
movement” could rapidly encompass every
consumer into a formal customer not
adult in the world. only accords dignity and respect but
The social impact also promises to be pro-
also opens the door to formal savings
found. BDSC offerings can spur the entry and insurance services, as well as
of many millions of aspiring middle-class
financial management tools
consumers into the formal credit system. The
conversion of a formerly “invisible” consumer
into a “visible” and formal consumer not only accords dignity and respect but also opens the door to formal
savings and insurance services, as well as financial management tools. It also allows low- and middle-income
households to better capture opportunities and manage economic shocks.6 Indeed, the advent and rise of BDSC
services serves to strengthen the continued growth story of the world’s emerging markets—a mega economic
trend of the 21st century.
In the rest of this section, we explore some forces that are shaping the digital revolution in financial services, and
delve into the opportunities and challenges facing the cohort of promising BDSC firms. In the two sections that
follow, we turn to the users of BDSC services themselves, sharing the results of our in-depth research into “early
adopters” in Kenya and Colombia. We report on the motivations, behaviors, and experiences of these aspiring and
middle-class consumers who are already taking advantage of these new offerings—and explore their views on two
salient issues facing this field in the future: privacy and trust. Finally, in the last section, we offer parting thoughts
on what each of several key stakeholder groups—innovators, financial services providers, “data owners”, and
regulators—must do in order to optimize the benefits and efficiencies, and also guard against the challenges, of
this rapidly evolving but early field.
4
Big Data, Small Credit is at the Intersection
of Three Significant Global Trends
Rapidly escalating processing power
that matches smart algorithms to
massive data streams
Figure 1
5
THE DIGITAL REVOLUTION AND FINANCIAL SERVICES
Historically, there has been a marked disconnect between what developing-market consumers need and the
relatively limited set of products and services that are available to them from formal financial services firms. Under
the old dynamics and economics of banking—the high costs of serving these consumers via traditional models,
the relatively small transaction values involved, and the difficulty of knowing and assessing many of these
consumers through traditional means (e.g., credit bureaus)—very few of these products and services have been
appropriate, affordable, or accessible to this huge but underserved market.
The digital revolution is rapidly changing these dynamics and creating new opportunities to address many of these
longstanding barriers in financial services. Indeed, an additional billion “off the grid” individuals are expected to
come online in 2017 alone.7 Meanwhile, mobile phone penetration is skyrocketing: GSMA estimates that more than
six in 10 global smartphone connections will come from developing regions within five years.8 The advent of the
low-cost Android smartphone is expected to further change the dynamics of financial access; it should also
expand the variety of services that individuals demand.9 Since 2010, consumers in India have been able to
download and listen to a Bollywood film song for Rs1 (or just over one U.S. cent).10 Reports show that, in countries
like the Philippines, Thailand, and Indonesia, social networks and, in particular, Facebook accessed via mobile
phones are the very definition of the Internet.11
As a result, the amount of digital information
generated by and about consumers on a daily
Together, these consumer, data, and
basis has exploded—creating a steep rise in
digital revolution trends are helping to available “big data,” a trend that has been well
change the landscape of inclusion and documented. CSC estimates that in 2020,
data production will reach 35 zettabytes—or
reach, offering the promise that billions
35 billion terabytes of data—which is 44 times
of previously “invisible” consumers can greater than the data produced in 2009.12
be “visible” for the first time 13
Seventy percent of this data is being gener-
ated by individuals, not companies—and not
just in the developed world, but everywhere.
In turn, processing power is also increasing exponentially. Smartphones now possess more processing power than
what was at NASA’s disposal when they sent the first men to the moon. As a result, companies across industries
as diverse as financial services, consumer packaged goods, and healthcare are racing to develop advanced
analytics capable of finding patterns and structure within the mass of digitized data being produced—and using
what they learn to create and tailor new offerings for consumers and markets.
Together, these consumer, data, and digital revolution trends are helping to change the landscape of inclusion and
reach, offering the promise that billions of previously credit “invisible” consumers can be “visible” for the first time.13
The nascent BDSC field sits at the intersection of the explosion of digital data, the rapid development of analytics
capable of mining this data for meaningful insight, and the aspiring middle-class consumers who are becoming
digital and discoverable for the first time. As this intersection, new opportunities and new consumer value propo-
sitions are being created that have the potential to enable formal financial services to viably reach and better serve
these suddenly “visible” and marketable consumer segments.
6
THE BDSC ADVANTAGE: LOWER COST, BETTER REACH
Extending loans in underserved markets has always been challenging. In the absence of easy-to-access
information about the creditworthiness of unbanked or underserved individuals, lenders have had to go to great
lengths to assess the likelihood of repayment. Indeed, the difficulty in collecting adequate predictive data on
underserved consumers in a cost-effective manner has plagued credit providers for decades.
That is now changing. Already, there is great activity in the BDSC space. A recent report by the Center for
Financial Services Innovation (CFSI) identified more than 40 companies in several countries that use big data
analytics or underwrite based on credit scorecards using nontraditional data, while a Consultative Group to
Assist the Poor (CGAP) report put the number that operate in the emerging markets at 36.14
Recent research suggests that using advanced analytics and nontraditional, large data sets to assess credit
worthiness offers a substantial cost advantage when it comes to providing credit in emerging markets. In countries
where traditional data doesn’t exist for the majority of consumers, it can dramatically lower the cost of identifying,
assessing, and reaching underserved consumers. CGAP, building on McKinsey and Co. analyses, estimates that
digital data analytics can reduce the marginal costs of providing a $200 loan in Tanzania by more than 40 per-
cent.15 Assuming such cost savings are reflected in their offerings, these economics should further accelerate the
affordability and uptake of the products and services offered by these companies.
It should also open up large and newly viable consumer segments, and make smaller loan ticket sizes profitable.
All told, our estimates suggest that the BDSC market approach could reach between 625 million and 1 billion
emerging-market consumers who are currently unable to access formal credit.16
THE ABCDs OF EARLY BUSINESS MODELS:
ALGORITHMS, BIG DATA, CONSUMERS, AND DISTRIBUTION
While the BDSC field has the potential to significantly alter the economics of credit in emerging markets, its
success will depend in large part on surmounting a series of challenges. Some of these challenges are integral to
BDSC firms’ innovation positioning; for instance, proving their algorithms’ predictive ability, which in turn depends
on their data sources, is clearly a make-or-break proposition. Some are part of their competitive positioning: how
successful they are at navigating the often-long B2B sales cycles and convincing incumbents to provide access
to their data and consumers. Still others are part of their strategic decision-making but lay mostly out of their
control—for example, the speed of digital penetration of country markets and their regulatory evolution. Below,
we identify the “ABCDs” of the BDSC field: four factors that we believe will determine the success or failure of
firms and their business models in this market.
a2 +
ba= Algorithms
As one might expect, there is still substantial experimentation with data algorithms, the key business ingredient
that provides accurate predictive value on segmentation and repayment. Although multiple variables can yield
signals, it can be hard to parse the contribution of each variable to the end score. Firms like Cignifi use algorithms
based on anonymized mobile phone call data records (CDRs) and SMS patterns. A recent case study of Airtel and
HFC Bank in Ghana suggested “Airtel customers using SMS heavily tended to have more frequent, higher-value
transactions in their HFC accounts; the value of monthly banking transactions was three times higher for those
in the top quartile than for those in the bottom. This points to the potential value of SMS usage as a proxy for an
7
individual’s discretionary income, as well as a powerful marketing avenue through which to reach high-utilization
customers.”17
Already, like Cignifi, many of these firms are developing the ability to assess the creditworthiness of previously
unbanked consumers at a significantly lower cost compared to the brick-and-mortar banking industry cost.
Several firms have suggested that their algorithms now outperform traditional credit scoring algorithms (such
as FICO scores) in the markets they target, even though FICO and similar industry incumbent scores have had
decades of refinement. As these algorithms continue to be tested across markets, segments, and services, their
quality and accuracy will likely continue to improve—and continue to bring down the costs of identifying and as-
sessing good credit risks throughout the developing world.
Moreover, in the event they are used as supplementary scores, these new credit scoring approaches add new
dimensions—such as social network behavior, mobile phone usage patterns, and even ““gamified”” experiments—
to existing scores to reveal more sophisticated consumer behavior and preference patterns. Already, in addition to
de novo algorithms, firms like Lenddo are combining their scores with traditional credit bureau scoring to unlock
more accuracy and create more powerful credit bureaus as a result. Nonetheless, firms will continue to experiment
and refine; no single definitive standard has emerged yet.
Big Data
Our analysis of firms suggests that they are working with varied data sources to do their scoring (see Figure 2).
Some of these sources are readily available in the public domain. Others require partnerships with telcos (as in the
case of call data records or CDRs) or with banks (when using the latter’s proprietary data sources) to refine and
scale their consumer segmentation and scoring approaches. Still other innovative providers are building their data
libraries directly from consumers—like Entrepreneurial Finance Lab (EFL), which aims to build a quasi-big data
repository by aggregating consumers’ psychometric test results. It is not yet clear from which of these sources
the strongest signals of predictive value will emerge—but the promise inherent in these varied sources of data and
their combinatory power is considerable. Unlike credit bureaus, which have longstanding data access agreements
across the financial sector and elsewhere, BDSC firms face uncertainty about the continued availability of the
data required to power their algorithms. This factor might force these firms to use primarily proprietary data from
companies that currently own this data, or rely on data directly provided by consumers. It also suggests that BDSC
companies can find a competitive advantage in reliably acquiring this data, and having the technology to process
and clean (what is often messy) data at scale.
It is also possible, even likely, that entirely new considerations—such as new forms of nontraditional data not yet
in existence—will emerge. The power of new sources of data to yield predictive signals is the very platform upon
which this space is built. Which kinds of data will prove the most predictive when it comes to evaluating
creditworthiness, and which will remain at best supplementary, remains to be seen.
Consumers
To be sure, BDSC offerings are still very young—most are less than five years old—though many of the companies
providing these services have experienced rapid growth in a short period of time. M-Shwari, for example, reached
7.2 million unique consumers with its bank-linked saving accounts within 24 months of its launch, which allowed it
to disburse 20.6 million cumulative loans (with a 2.2 percent default)18 Lenddo—a service built on the notion that a
person’s social media activity can be a high-quality proxy for their creditworthiness—assessed hundreds of
8
Multiple Data Types
Call Data BIG DATA Public
Patterns Databases
Web Browsing Government
History Records
Social Location
Media Activity Data
Mobile Money
Transaction Data
Utility Payment ID, Income,
Records Residence Proof
Behavioral Psychometric
Analytics Data
NOT-SO-BIG
DATA
Figure 2
9
thousands of loan applications throughout the Philippines, Mexico, and Colombia during its first four years
of operation.19
All of the BDSC providers we met with in the course of developing this report confirm that there is clear consumer
demand for unsecured, formally provided, reasonably priced credit for short-term purposes. The main questions
are, which previously unbanked consumers will prove to be early adopters, which will turn out to be difficult
segments to reach and score, and how will their opinions and behavior change over time? Given that these are
new consumer segments for formal lending, much remains to be learned about them. Indeed, that is the focus of
the following two sections of this report.
DISTRIBUTION
There is still no consensus on an optimum model for how young BDSC firms can reach these millions of new
consumers. As a consequence, BDSC business models vary widely: some BDSC firms engage directly with
consumers, others rely on access provided by incumbents, and a few have tried both combinations. These routes
can be circuitous: for instance, Lenddo relies on social media companies’ APIs to gain access to the consumer,
although they must get consumers to opt in.20 Currently, some BDSC firms are positioned as direct lenders to
consumers, and others are structured as pure data science companies offering algorithm-driven services to
telecommunications companies, banks, insurance companies, microfinance organizations, and other incumbents.
Solving the distribution issue will likely take time, as well as continued experimentation with both partnerships and
direct lending. However, the desire to “own” the consumer relationship will operate in tension with the need to
partner with, or sell to, bigger players –particularly since the latter option may accelerate impact at scale.
AN EVOLVING FIELD
How this field will evolve across these factors cannot yet be known. Yet, BDSC firms are certain to continue to
experiment with, tweak, or even pivot their business models many times over to respond to such challenges and
sustain competitive advantage. The winners will create better algorithms and products, more engaging consumer
value propositions, and shift to meet (or be ahead of) evolving regulations and market dynamics. We should also,
of course, expect some of these models to fail, which is a natural part of any market evolution.
Meanwhile, BDSC firms aren’t the only ones evolving. The needs and wants of BDSC consumers are also
changing—and BDSC firms will need to be willing and able to adapt to ever-shifting consumer behavior and
demand. Current consumers are already responding to and interacting with BDSC offerings in unanticipated
ways. And not all consumers are happy with their loan experience. For BDSC to fulfill its promise of serving new
“formal” consumer segments, it is critical to listen to the voice of the early adopter. This is what we turn to in the
next section.
10SE C TIO N 2
BDSC EARLY ADOPTERS:
WHY AND HOW DO THEY BORROW?
“Rose,” a 46-year-old shop owner in Nairobi, Kenya, lives in a household of six and did not complete her
high school education. But she earns roughly 15,000 KSH ($167) per month, which is 60 percent more
than the average Kenyan. And she has been saving money in a bank account for the long term, hoping
someday to open a chain of shops.
In the past, Rose has borrowed money from microfinance institutions, from a Chama (an informal
cooperative that pools and groups savings), and from family and friends. Her loans have ranged from 1,000
KSH ($11) to 30,000 KSH ($335), and most of that money has been used for her business. Rose thinks
carefully about who she borrows from, however. “I only borrow from people I know,” she said. “We can
even talk in Swahili or Luo [and they] live near me.” She considers sharing certain private information a
normal part of taking out a loan. But there are some categories she would never share, including
information about her family’s medical history, the websites she visits, and her emails.
11Rose’s preferred and trusted lender has historically been the Chama, in part because she is very familiar
with the group’s activities. Apart from the Chama, Rose trusts state banks and mobile money services. In
general, she only trusts financial organizations “if they tell me about all the charges in advance.”
Rose learned about M-Shwari from a TV advertisement and signed up—but she did not apply for an
M-Shwari loan until she had an emergency. Her mother died and she needed money to pay for the funeral.
“I first turned to my brothers and sisters, but nobody seemed to have money,” she explained.
“So I asked for 7,000 KSH ($80) from M-Shwari, was granted the loan, and received money within hours.”
Rose told her daughter about M-Shwari as well, and it is now their preferred lending service.
Rose said M-Shwari is the best lender in an emergency situation because it is simple and fast, unlike
formal financial institutions. “Last holiday when I was on safari and ran out of gas, I borrowed 4,500 KSH
($50) from M-Shwari and fueled the car…as the bank was some distance away.” Her recent loans have
all been relatively small: 1,000 to 5,000 KSH ($12 to $56). Rose describes the M-Shwari experience as
“smooth” because she has received all the money she wanted to borrow. She is happy with its services.
THE VOICE OF THE CONSUMER
Rose is an example of a consumer once thought to be “invisible” to financial services providers who is now
“digital and discoverable” thanks to BDSC. Indeed, she is a BDSC early adopter—one of hundreds of thousands
of individuals currently testing out these new products and services in emerging markets.
Rose is fairly typical of the BDSC early adopters we surveyed. She’s urban, and also more educated and more
financially savvy than the average consumer in her market. She owns a smartphone and uses it for calls and
texts—but also for social media, financial transactions, and other applications. She has a comfortable income but
encounters occasional, abrupt financial needs that require loans. She is aware of her borrowing options and seeks
out the most appropriate and convenient services. In other words, BDSC services are not helping Rose become
“financially included” as much as “financially better served.”
Despite growing literature about the rising BDSC field, very little has been captured about the individuals who use
these services: aspiring middle-class consumers in developing markets who are now gaining access to needed
short-term credit in new ways and more easily than ever before. In order to better understand this large and
emerging consumer cohort, we commissioned in-depth research into their behaviors and preferences.21 This
research featured individual interviews with more than 300 adults living in major cities in Kenya (a low-income
country) and Colombia (a middle-income country). Seven in 10 were current users of BDSC services; the others
were aware of and had access to these services but had not yet pursued them.
This section offers a profile of these current and potential users of BDSC services in these two emerging markets,
illustrating a broad range of use cases, borrowing behavior, consumer preferences, and concerns. Understanding
the needs and motivations of such “early adopters” in markets where BDSC services are gaining traction is key to
the design of tailored, cheaper, and more appropriate products and services—as well as to future business
models. These early adopters alone already represent a very attractive and underserved consumer segment,
and a sizable opportunity.
12And this is only the beginning. If BDSC models
become true mainstream models, eventually there will
be more and more consumers unlike Rose in the mix,
as the services expand to become more mass market
and include additional, low-middle income consumer
segments. In other words, widespread adoption of
BDSC models should bring into the fold many
consumers who are new to formal unsecured
credit—consumers who have thus far remained
“invisible” despite the rapid advances in financial
technology and services. Their opinions have not yet
been taken into account in building out and extending
financial services ecosystems. As BDSC products
gain traction, it is our hope that many millions of these
consumers will also be brought into the fold.
EARLY ADOPTERS: YOUNG, Who We Spoke to—and Why
SAVVY, AND EDUCATED Our consumer research was driven by the belief that
Not surprisingly, given the digital nature of most of the critical insights into the present and future of the BDSC
field will come from listening to the voices and the
products available, the vast majority of interviewees
perspectives of the individuals using these services
were younger and more educated than the national themselves. These voices are critical for understanding:
average (see Figure 4). For instance, 70 percent of the
• How these new services fit into consumers’ financial
Kenyans we interviewed were 34 years of age or
behaviors and credit aspirations
under, and 91 percent had at least a secondary
diploma or vocational training. In Colombia, those • The BDSC value proposition for users, and how they
plan to use these funds
percentages were 68 and 97 percent, respectively.
In both countries, nine out of 10 interviewees were • Their concerns about multiple issues related to BDSC
“comfortable” with their income. In Colombia, 84 services, from service quality and convenience to trust
and privacy
percent were able to save or buy expensive goods
after meeting their basic needs; in Kenya, it was 77 Bringing these perspectives to light at an early stage
percent.22 In other words, these early adopters of the BDSC field’s evolution could help shape the
commercial evolution of the emergent BDSC space
already represent an attractive consumer segment
and the policies established to support it. It might also
for any lender. contribute to the creation of a balanced architecture
that encourages risk, innovation, and entrepreneurship,
Interviewees were frequently self-employed but also while also valuing privacy, ensuring protection, and
more stably employed compared to the general enhancing trust.
population and more savvy with respect to both finan-
To be sure, the consumers we studied are all early
cial service providers and mobile smartphone use: adopters, and we have little doubt that BDSC user
50 percent of Kenyans interviewees (versus 6 percent dynamics will evolve over time. Additionally, our sample
in Kenya as a whole) and 78 percent of Colombians cannot be considered nationally representative of either
Kenya or Colombia. However, our research revealed
(versus 30 percent) owned a smartphone. They used
valuable information about consumer perspectives and
these smartphones for a wide variety of transactions, early experiences that should nonetheless prove highly
from phone calls to social media to financial valuable to those invested in the evolution of the field.
13Provider Services
Figure 3
transactions. These findings are consistent with what one would expect among early adopters of a product that is
both new and digital in nature.
Indeed, the digital proficiency of these early adopters was notable. In both countries, the majority of interviewees
reported conducting three to four different kinds of operations on their smartphones each day (e.g., browsing the
Internet, using social networking sites, and making financial transactions). They also demonstrated comfort with
advanced services, including downloading applications, requesting on-demand services, and navigating menus
and choice architecture.
Of particular note in Kenya was the pervasiveness of mobile payments and financial transaction applications
(which is not yet the case in Colombia). In Kenya, 50 percent of respondents accessed digital media exclusively
on their smartphones. In Colombia, access was spread more evenly between phone and computer. In both
countries, social media formed the bulk of new media use: 64 percent of the Kenyans and 46 percent of the
Colombians interviewed indicated this as their favorite online activity.23 Seven out of 10 interviewees had an
account with at least one social networking site, and Facebook and WhatsApp were the leaders among these
sites in both countries.24
14Figure 4
15
Figure 4In other words, these BDSC early adopters tend to be younger, more educated, banked, smartphone savvy, stably
employed or income-generating, and more risk-tolerant that the average consumer.
This dynamic is an important one, with precedent in recent financial services mass-market adoption. For instance,
although today large swathes of M-PESA consumers come from low-income, rural, and unbanked households,
they were not among its early adopters. Indeed, Safaricom first marketed M-PESA to the young, urban middle
class. Consequently, when M-PESA was not yet two years old, its early adopters were largely tech-savvy, banked,
educated consumers with significant discretionary income. Critical to the success of M-PESA’s mass-market
adoption were these early adopters, who forged an aspirational brand and achieved critical mass to fuel
network effects.25
BORROWING BEHAVIOR
Borrowing money was a critical aspect of the financial lives and financial behaviors of interviewees in both coun-
tries. At the moment, these early adopters are using BDSC products mainly as additional or supplementary tools
for transactions where existing (formal and informal) tools fail, so they have quickly found value in the marketplace.
These early adopters were overwhelmingly banked (80 percent in Kenya, 99 percent in Colombia) and already had
relationships with formal financial institutions of some sort. In Kenya, most had mobile money accounts.26 They
used banks for large transactions and long-term savings, and mobile money for P2P transfers, short-term savings,
and everyday payments and purchases.27
Even though Kenyan interviewees charted as relatively affluent and comfortable, one common aspect across
both Kenyan and Colombian interviewees was that they had all applied for loans multiple times and taken out
loans when they were granted. Nearly 60 percent of their borrowing was originally driven by the need to cover
unforeseen expenditures and debt repayment. However, as Figure 5 shows, once they had borrowed, they spent
nearly 80 percent of their loans on such needs. Kenyan consumers, in particular, still faced multiple occasions
where they had short-term or emergency borrowing needs—the number one reason for borrowing among
interviewees—and they viewed BDSC services like M-Shwari as far more convenient and responsive to those
needs than any other option in the market.
Even though Kenyan early adopters were wealthier and more savvy than average Kenyan consumers, and 80
percent had bank accounts, they still faced day-to-day cash flow and borrowing needs that these banks could
not fulfill, signaling a volatility in incomes even at higher income segments.28 This was particularly so for the women
who participated in our survey.
In Colombia, although there was less emergency borrowing, expenditures to invest in consumers’ businesses or
homes (e.g., bicycles, car payments) or even routine expenditures (e.g., education, school fees) accounted for 70
percent of borrowing. In other words, even among this segment, and even in Colombia, borrowing is a key tool in
managing and offsetting volatility—rendering the ability to access quick, timely, and convenient services as very
important.29
It is also worth highlighting that the ecosystem for credit and banking services is very different in Kenya than it
is in Colombia. As Figure 6 shows, though informal lenders—including family members, friends, neighbors, and
co-ops— dominate the Kenyan lending market, BDSC and mobile money services are the second most common
source of loans. Indeed, three out of four interviewees borrowed at least part of the time via mobile channels. In
16Top Uses for Loans
Kenya Colombia
61% Emergencies 5%
Small Business
39% Expenses or 69%
Home Improvement
Routine
33% 12%
Expenses
19% Bill Payment 12%
19% Education 21%
Figure 5
17Figure 6
Colombia, 85 percent had applied for loans from banks, reflecting a stronger preference for and trust in the formal
financial system.30 Still, nearly six in 10 preferred to rely on other people, such as friends or family, for their loans.
BDSC services competed well in these ecosystems for small loans, because of their faster and more simplified
loan processes, and relatively low interest rates. Indeed, in Colombia, the top reason participants did not borrow
from banks for shorter-term credit was high interest rates, cited by more than 20 percent of respondents.
BDSC is in its infancy in Colombia, and growing somewhat faster in Kenya. But in both markets, early adopters are
driving uptake and growth. The demand for BDSC services as a substitute for existing services, both formal and
informal, is strong among these segments. They find the convenience, the speed, and the service attractive, and
they tailor their BDSC usage to small transactions. In Kenya, BDSC typically functions as the first formal lender,
replacing or supplementing informal sources. In Colombia, BDSC is replacing either the private bank or the
informal lender.
Overall, consumers indicate strongly that BDSC offers them a convenient, formal alternative to the informal sources
they usually borrow from—and so they find the value proposition attractive.
However, whether BDSC services will deepen and expand in these markets will depend in large part on whether
consumers grow to trust these services. In the long run, BDSC success will to a large extent be determined by
whether consumers gain confidence in the way that BDSC lenders use their private information.
18SE C TIO N 3
BDSC EARLY ADOPTERS:
THEIR EVOLVING TRUST CALCULUS
“Erika,” a 26-year-old aspiring marketing executive in Bogota, Colombia, works full time as an office
assistant while trying to finish her college degree. She lives in a household of four and earns about 800,000
COP ($385) a month. She has had a savings account with Banco Colpatria for 10 years but uses it only for
basic transactions.
In the past, Erika has applied for loans with financial institutions and also asked friends and family for
money. Her prior loans have ranged from 1,000,000 COP ($480) to 4,000,000 COP ($1,924). When applying
for a loan, she wants an institution that is recognized in the market, has lent money to other people she
knows, and has a rapid turnaround: “I’m looking for transparency in financial institutions, where they
provide more information about their portfolio on the Internet, where it can be easily accessed.”
Erika believes that sharing private information—including her email address, phone number, employment
details, and financial history—is a normal part of the loan application process. She’s willing to disclose
information she would normally only share with her husband, such as social media information, if it speeds
up the loan process, which is her top selection criteria when it comes to choosing a loan provider.
19In recent years, the related issues of privacy and consumer trust have risen to the forefront of public debate, largely
due to the unprecedented access that governments, companies, and civil society now have to the digital footprints
that billions of people create when they transact and communicate online. While the ability to process and analyze
personal data has the potential to unlock tremendous insight into consumers around the world, as we have argued
earlier in this report, it also continues to trigger concerns about privacy. Where should the line be drawn when it
comes to the use of personal data?31 Does that line differ when it is for marketing and advertising versus collecting
information on opt-in services? What rights do individuals have when it comes to controlling and protecting their
personal information—not just from people who mean harm, but also from people who might mean good?
As with other sectors like healthcare, education, and telecommunications, many of the financial services
available today—from online lending to mobile payments—involve frequent exchanges of confidential consumer
information over digital pathways. Much of this information is stored, creating significant repositories of data on
billions of individuals.
Despite this large digital exhaust trail, currently, those of us in the financial inclusion field know almost nothing
about the way potential emerging-market consumers think about confidential information, privacy, and trust—
particularly in the realm of these digital financial services. Are they aware that they are providing personal data in
exchange for services? To what extent—if at all—do privacy considerations affect their decision-making or
willingness to engage with providers who use these new forms of scoring? Do consumers perceive a tradeoff
between participation and protection in the formal financial system? And what related trust issues do BDSC
services pose for such consumers?
There has not been a systematic effort to understand emerging-market consumers’ attitudes toward privacy, trust,
and the tradeoffs they perceive or are willing to make in order to secure BDSC—or indeed other financial—services.32
But these issues are rising on the agenda of regulators, consumer protection advocates, BDSC providers, their
business partners and investors, and many others in this ecosystem.33 Understanding consumer perspectives on
data confidentiality, what components of their data should remain private (and over what time horizon), and what
level of privacy they are willing to give up could have an impact on BDSC market regulation, usage protocols, and
ultimately the sustainability of BDSC business models themselves.34
TRUST, PRIVACY, AND CREDIT: THE BOTTOM LINE
Our consumer research in Kenya and Colombia strongly suggests that early adopters are well aware of the issues
and tradeoffs around confidential information, trust, and privacy that they face when engaging with financial service
providers in general, and BDSC offers in particular. Three key insights stood out from the perspectives of the BDSC
consumers we interviewed in Colombia and Kenya, each critical to how they engage with these issues.35
1. Consumers think of privacy on a continuum, with some information ranking as “more private” and some
as “less private.”
As Figure 7 shows, the consumers we interviewed ranked their financial history and income level, the content of
online/mobile phone communication, medical histories, and family particulars as their most private information.
They reported that they typically would only share this information outside their immediate, trusted circle of family
and friends in the event of an emergency. Conversely, almost nine in 10 consumers across both countries had little
or no concern about sharing their national identity documents in exchange for financial services, even though five
in 10 ranked national IDs as private information.
20What Data Consumers Consider Private
Figure 7
21As the ID example indicates, both Kenyan and Colombian consumers share nuanced views about what they do or
do not consider private. In other words, they understand well the value of their personal data and its requirements
for different services (e.g., voter registration, “know your customer”). This should underscore their ability to exercise
agency in financial transactions, especially in instances when they are making tradeoffs and decisions about sharing
and using this data.
2. Most consumers are willing to exchange their data—even their most closely held, confidential data—
in exchange for bigger or better loans.
As Figure 8 shows, the consumers we interviewed were willing to share multiple kinds of data, even those items
they ranked as more private—including mobile phone use, social networking use, general Internet use, and bank
account activity—to improve their chances of getting a loan or to increase the size of that loan. In the case of
Colombian consumers who borrowed from Lenddo, we found that roughly 70 percent were willing to share infor-
mation on their social media activity and web browsing history with Lenddo to improve their chances of getting
a loan or getting a bigger loan.36 This finding is mirrored in the CGAP work with First Access in Tanzania: “Many
consumers noted that the need for a loan would supersede concerns for privacy, making clear why it is important
for providers to develop standards for informed consent and data protection up front in product development.”37
In other words, for these emerging-market
consumers, the availability of formal, convenient,
unsecured, short-term, small-dollar credit was more
important than privacy—which makes the BDSC Meaningful consumer trust in
value proposition very compelling. In and of itself, this
financial institutions in low- and
consumer view of the tradeoff between privacy and
credit is a significant finding, and comes at a key time middle-income markets takes
in the development of this industry. Yet it is not time to build and effort to sustain
entirely surprising, given that BDSC early adopters
are more familiar with financial institutions and already
expect that they will need to provide personal data
in order to secure credit. Indeed, in Kenya, only 5
percent of all participants interviewed said they felt uncomfortable when a financial institution asked them to share
their private information as part of their “know-your-customer” or KYC process.
It is worth noting, though, that as the market deepens and broadens, we would expect the current calculus to
evolve for both established and new consumers. When we asked non-BDSC users in Colombia what private
information they would divulge in exchange for a better chance to get a loan, or to get a bigger loan, there was
markedly less willingness to share social networking data (although still a majority agreed). There was also some-
what lower willingness to share Internet data usage, and, conversely, much higher willingness to share information
on bank account use. It is likely that certain segments will become more concerned about transparency, consumer
protection guidelines, privacy protocols, and opt-out models that circumvent informed consent. While we cannot
predict how these dynamics will evolve, these are important issues that warrant thoughtful consideration and
monitoring over time (see Section 4).
22Figure 8 23 Figure 8
Figure 9
3. Consumers worry about the security and integrity of their data. They know what would constitute a
violation of their trust (e.g., identity theft, selling private data to a third-party, or giving it to government
tax authorities) and have strong ideas about how BDSC firms might work to enhance their trust.
Consumer concerns about privacy—as well as their willingness to share confidential information for better financial
services—are intimately connected with the issue of trust. It is clear from our interviews that meaningful consumer
trust in financial institutions in low- and middle-income markets takes time to build and effort to sustain. Key
elements for building that trust include transparency; a strong reputation in the marketplace; a record of honoring
previous commitments; endorsement from respected persons; and responsive, rewarding consumer relationships.
Strikingly, as Figure 9 shows, a single bad experience or erroneous transaction can undo much of the trust that
consumers are willing to place in a financial institution. Asked specifically about privacy issues, consumers’
willingness to share private data for a loan did not diminish other important considerations. In Kenya, about 41
percent of those we interviewed were concerned about their information being given to the government revenue
authorities. In Colombia, 87 percent worry their information would be accessed by criminals (and were far less
concerned about tax authorities). But these considerations did not overshadow their willingness to share sensitive
personal data with lenders in order to access credit.
While these concerns stem mainly from previous interactions with a variety of financial service providers, they
nevertheless have important ramifications for new BDSC entrants with low brand recognition seeking to enter the
market with alternative credit-scoring-based offers. Again, we expect these attitudes to change and evolve as
consumers become more entrenched and experienced in the formal financial system, and as more financial
services become available to them.38
24LOOKING AHEAD
In the world of fast-paced and evolving digital financial services, where confidential information is frequently
and quickly transacted, the assurance of privacy and the honoring of transparency are critical aspects of the
relationship framework between clients and institutions, especially by newer firms that lack the brand heft
of incumbents. Indeed, these two features are foundational, underpinning both trust and sustainability.
And still, in this current moment, there is little standardized policy governing who owns what data, and how it is
used, stored, and managed.
This finding carries important implications for regu-
lators and firms serving these consumer segments
—and raises important questions about how BDSC
BDSC models should be designed
companies will navigate the issues of trust and
privacy in these and other markets. How will they with these consumer concerns and
make certain that consumers understand their tradeoffs in mind, not just for one
service usage agreements and products? How
transaction but for as long as these
should they tackle consumer concerns about private
information and assure them that their data is safe? relationships are held
Is informed consent necessary, or even enough?
And how will they balance these requirements with
the ability to innovate to serve?
One likely implication of these findings is that BDSC models should be designed with these consumer concerns
and tradeoffs in mind, not just for one transaction but for as long as these relationships are held. The field will likely
evolve, at least in some countries, to require more transparency over time—giving consumers greater control over
their user agreements, their ability to opt in and out, how third parties can use their data, and their options for
addressing grievances should these assurances not be met. It may well be that there is a comparative advantage
for companies that design their models with consumer control, transparency, and disclosure incorporated as key
principles from the outset.
25SE C TIO N 4
DELIVERING ON BDSC’S PROMISE
Early adopters of BDSC services, for whom borrowing will continue to be a major part of their financial lives, say it
offers them a compelling, convenient, and affordable borrowing alternative. Consequently, the demand for BDSC
services as a substitute for or enhancement of existing services—informal and formal—is strong, particularly for
small and high-frequency transactions. This represents a massive opportunity with huge potential upside—both
social and commercial—if harnessed well. Indeed, this is the promise of Big Data, Small Credit.
However, these are still early days. As with all such disruptive spaces, we can expect much to change in the overall
market landscape, as well as in the relationship dynamics between innovators, incumbents, and rule-makers. How
will key players in the system receive the BDSC value proposition? Will regulators allow these BDSC models to
continue to experiment and innovate? Will data owners and financial service providers, who currently own both
data and consumer relationships, look to capture the opportunity themselves, partner with BDSC firms, or work to
block these innovators? How will the field secure and maintain a trust-enhancing architecture that sets the platform
for consumer financial wellbeing?
It will be awhile before we know the answers to many of these questions. Important execution challenges have to
be overcome before BDSC becomes mainstream: as seasoned investors, we would be naive to expect otherwise.
26A CALL TO ACTION
Innovators • Focus on getting the “ABCD” elements of the business model right:
Algorithms, Big Data, Consumers, and Distribution
• Treat client data with integrity and use it only for its
stated purpose
• Give clients the option to opt out of the use of their data
• Invest in building secure technology platforms to protect any sensi-
tive client data that you access
Financial Services • Engage with BDSC firms to reach new, profitable consumer
Providers segments
• Pilot new approaches quickly to stay ahead of competitors aiming
at the same segments
• Do not fear the disruption to traditional bank underwriting models
that these new innovations represent
• Leverage the trust and brand advantage but compete on
user experience
Data Owners • Share data with BDSC firms in order to build the broader
ecosystem
• Give consumers greater control of their data to allow them to make
informed choices
• Partner with the innovators to open up new lines of business
Policymakers • Allow space for innovation to occur before regulating too
heavily—e.g., follow U.S. example and issue “no action” letters
on key innovations
• Monitor closely for traction and unintended consequences
• Above all, provide clarity on intentions and rules
• Focus on key issues such as the systemic risk posed by easy
access to credit that BDSC might facilitate, but not at the expense
of allowing for early experimentation
27It has taken decades to bolster the accuracy and acceptance of traditional credit ratings; it took many years in the
U.S. for FICO to become the mainstream score. Similarly, we must be patient as BDSC innovators figure out what
data sources and analytics provide the best predictors of creditworthiness, and how best to deliver formal services
to consumers.
To deliver on the BDSC promise and to capitalize on the significant opportunity it can afford will require concerted
action from a key set of stakeholders: financial service providers (e.g., banks, MNOs), data owners (often technology
companies and data aggregators), innovative entrepreneurs, and regulators. Below, we set out the key consider-
ations and a call to action for each of these principal groups.
INNOVATORS:
IMPROVE BUSINESS AND PARTNERSHIP MODELS
We have yet to see one business model eclipse the BDSC landscape in the way that Lending Club and Prosper
have done in U.S. marketplace lending, no doubt helped by Lending Club’s highly visible, multi-billion dollar IPO.
There are still many active business models in this space, and we expect that diversity to continue for some time.
Indeed, many of the factors driving this diversity are internal, revolving around the continued evolution and
refinement of the algorithms and partnership models that are foundational to these businesses.
Can these startups create compelling use cases to convince incumbent data holders to sustainably provide
access to the consumer data needed to scale new scoring methods and pursue new segments and markets?
Will public-domain data sources be most valuable, or will more private, proprietary sources prove more useful?
Will innovators with direct-to-consumer models prevail,
or those that rely on B2B incumbent channels and
distribution? How important will it be for BDSC firms Innovators need to continue to
to own the consumer relationship? Which consumers
refine and demonstrate their
and segments will form the mass-market core of the
consumer base? And finally, will firms capitalize on the algorithms, adapt their business
competitive advantage likely to be gained from being models, and create compelling
transparent with consumers, secure with their data,
and profitable use cases
and cognizant of their privacy concerns?
There are external issues beyond BDSC firms’ immediate
control as well. Will the revenue generated from these consumers actually prove to be significant enough to
entice incumbent financial services providers to partner in the first place, if indeed the route to success goes
through B2B distribution partnerships? Will enabling policy accelerate the market adoption of the technology and
digital access required for BDSC assessment to work?
Innovators should focus on the ABCD factors outlined earlier in this report, and on building out key elements of the
“trust architecture” that must be created. Those that are effective at the latter are likely to build a loyal consumer
base and therefore benefit from those consumers’ repeat business. In addition, innovators need to continue to
refine and demonstrate their algorithms, adapt their business models, and create compelling and profitable use
cases that deliver quick wins for incumbent partners and for themselves.
28You can also read
